personalization-at-scale
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references an external Model Context Protocol (MCP) server endpoint (
https://mcp.manypi.com/mcp) associated with the skill's author. This endpoint is used to add research capabilities to the agent via a shell command. - [INDIRECT_PROMPT_INJECTION]: The skill outlines a process for ingesting external data from the live web (such as job postings, blog content, and customer reviews) into email templates. This creates a potential attack surface for indirect prompt injection if the source data contains malicious instructions.
- Ingestion points: Data is pulled from external sources into custom fields like
recent_signaland homepage content. - Boundary markers: Data is merged into templates using standard
{{variable}}delimiters. - Capability inventory: The skill uses the agent's natural language generation capabilities to transform researched facts into email copy.
- Sanitization: The skill focuses on 'data hygiene' for formatting purposes (e.g., casing and legal suffixes) but does not explicitly implement security-focused sanitization of the external content.
Audit Metadata